Life Science Journal 2014;11(6s)
http://www.lifesciencesite.com
Upgrading of Commercial Potassium Chloride by Crystallization: Study of Parameters Affecting the Process
Khaled Rawajfeh*, Thamar Al-Hunaidi, Motasem Saidan, Zayed Al-Hamamre
Chemical Engineering Department, Faculty of Engineering & Technology, The University of Jordan, Amman
11942, Jordan
[email protected]
Abstract: The production of high purity KCl from commercial potassium chloride (96% KCl) produced at the Arab
Potash Company in Jordan is investigated using Mixed-Suspension Mixed-Product-Removal crystallizer. The effect
of both the residence time and the level of supersaturation on the purity of the crystalline product and on the kinetic
parameters of crystallization were investigated. It was found that the purity of the produced KCl salt was adversely
affected by both parameters. Nucleation rate was found to increase as the level of supersaturation increases while the
growth rate was found to decrease. Both nucleation and growth rates were found to decrease as the residence time
increases.
[Rawajfeh K, Al-Hunaidi T, Saidan M, Al-Hamamre Z. Upgrading of Commercial Potassium Chloride by
Crystallization: Study of Parameters Affecting the Process. Life Sci J 2014;11(6s):8-15]. (ISSN:1097-8135).
http://www.lifesciencesite.com. 2
Keywords: KCl crystallization, ionic impurities, kinetic parameters, MSMPR crystallizer, purification, Jordan
and crystals [4]. Sayed and Larson studied the effect of
the addition of NaCl and MgCl2 on the crystallization
kinetics of KCl and on the crystal habit [5]. They found
that the presence of these salts in low concentrations
reduced markedly the tendency toward polycrystal
formation. They also found that the presence of Mg+2
ions reduced the amount of NaCl in KCl crystalline
product. Koenig and Emons and Koenig et. al. studied
the effect of NaCl impurity concentration on the
kinetics of KCl crystallization [6, 7]. They found that
both nucleation and growth rates decreased as NaCl
concentration increased. Akl et. al. studied the effect of
the addition of NaCl and MgCl2, separately and
combined, on the crystallization kinetics of KCl [8].
They developed correlations for growth and nucleation
rates and found that the addition of these salts
decreased the growth rate and accelerated the
nucleation rate. They also found that MgCl2 had more
pronounced effect on the crystallization kinetics than
NaCl. Rohani and NG studied the crystallization of
KCl from an aqueous solution co-saturated with NaCl
and KCl in the presence of small amounts of MgSO4,
CaSO4.2H2O, FeCl3.6H2O, CrNO3.9H2O and insoluble
silica powder in a 750 ml cooling batch crystallizer [9].
They investigated the effect of the admixtures on the
crystallization kinetics, crystal habit, and the roughness
of the faces of the crystals, the shape of the corners, the
polycrystal formation and the crystal yield. Dash et. al.
investigated the effect of magnesium and sulfate ionic
impurities on the crystallization of KCl using MSMPR
crystallizer [1]. Their results exhibited that Mg+2 did
not have any significant influence on the crystallization
of KCl; while the sulfate ion enhanced the aggregation
rate. The nucleation rate was slightly affected by the
presence of impurities. The role of the studied
1. Introduction
Crystallization process is one of the oldest unit
operations which has a widespread use in the
production of pure chemicals, pharmaceuticals,
fertilizers, etc.
Crystallization in the presence of impurities is
generally encountered in the chemical industry. For
example, in fertilizer industry, crystallization of KCl
from saturated solution is, in general, performed in the
presence of ionic impurities such as Na+, Ca+2, Cl-,
SO4-2. These impurities and many others, have great
effect on the crystallization kinetics, crystal habit,
purity and yield.
The effect of ionic impurities on the
crystallization process has been studied by many
investigators, and no general theory has been yet
formulated to explain their effects. Dash et. al.
indicated that the effects of impurities are due to their
ionic properties [1]. Bostaris et. al. studied the growth
of single seed crystals of KCl under constant conditions
of temperature, supersaturation, and impurity
concentration in the presence of Pb+2 ions [2]. They
suggested a mechanism of non-equilibrium capture of
the impurity and indicated to reject the first crystals
grown in order to produce KCl crystals free of Pb+2.
Sarig et. al. studied the effect of NaCl, MgCl2 and
AlCl3 on the crystallization of KCl in a batch cooling
crystallizer with either linear or non-linear profiles of
cooling [3]. No significant effect of these impurities on
the crystal habit, median crystal size, or coefficient of
variation was observed.
Two mechanisms on the influence of impurity on
crystallization were proposed by Mullin, the first one is
due to the electric field of the ions; while the second is
due to the chemical interaction between the impurities
8
Life Science Journal 2014;11(6s)
http://www.lifesciencesite.com
impurities upon KCl crystallization was found to be a
function of the chemical character of the type of
impurity. They were also able to identify two
mechanisms for size enlargement: the first was due to
crystal growth by solid deposition and the second was
due to aggregation which was found to be the
governing mechanism for producing large crystals.
In this work, MSMPR crystallizer is used to study
the effect of the degree of supersaturation and the
residence time on the crystallization of KCl from an
aqueous solution saturated with commercial grade KCl
salt produced by Arab Potash Company (APC). The
purity of the crystalline product and crystallization
kinetics are investigated.
size and it is determined by the interaction of the
nucleation and growth rates in a crystallizer.
An idealized crystallizer model called the Mixed Suspension Mixed-Product-Removal model (MSMPR)
is used as a basis for identifying the kinetic parameters,
namely, the nucleation and growth rates, and hence for
estimating the CSD of the crystallizer product. This
crystallizer is assumed to consist of a well agitated
vessel supplied with a solid-free feed solution at a
continuous steady rate from which a product stream is
removed at steady continuous rate having a CSD and
solids concentration exactly the same as that in the
crystallizer. Also in this crystallizer, which is operated
at steady state, the McCabe’s ΔL Law is assumed to
hold; that is the crystals of all sizes grow at the same
rate, ie crystal growth is independent of crystal size;
and also they have the same shape and no gross crystal
breakage occurs. The basic quantity in the theory of the
CSD is the population density n. Graphically, the
population density is the slope of the cumulative
numbers vs. size curve:
2. Theory
In common crystallization processes solid
particles are formed within a homogeneous saturated
liquid phase. This process is very important industrially
because it provides a practical method to produce pure
chemical materials from solutions of various salts.
The crystals are produced from supersaturated
solutions for which supersaturation is often expressed
as an equivalent temperature difference and it may be
generated by one at the following methods [10]:
a) Simple cooling and temperature reduction.
This method is used when the solubility of the solute
increases strongly with increase in temperature.
b) Evaporating a portion of the solvent. This
method is common when the solubility is relatively
independent of temperature.
c) Salting out by adding a third component
which acts physically with the original solvent to
produce a mixed solvent in which the solubility of the
solute is sharply reduced. This method is used when the
solubility is very high and neither of the above methods
is desirable.
d) Precipitation by adding a third component that
will react with the original solute to produce a new
insoluble solute. Using this method, which is used
when nearly complete precipitation is required, rapid
creation of very large supersaturation is possible.
Crystal formation is a two-step process [10]:
a) The birth of a new particle (nucleation) at a
microscopic size.
b) The growth to a macroscopic size.
The driving potential for both rates is
supersaturation. Nucleation is a consequence of rapid
fluctuations on a molecular scale in a homogeneous
phase that is in a state of metastable equilibrium.
Crystal growth is a diffusional process modified by the
effect of the solid surfaces on which the growth occurs.
A crystallization system can be fully described by
material and energy balances together with the Crystal
Size Distribution (CSD) of the crystallizer product. The
CSD is the relationship between crystal numbers and
(1)
where:
L: crystal size
V: volume of mother liquor in the magma. It is
considered to be constant since the MSMPR
crystallizer is operated at steady state.
N: number of crystals of size L and smaller in the
magma.
Applying the steady state population balance for
the size range L1 to L2 with population densities n1 and
n2 respectively, the balance is:
Number of crystals in = Number of crystals out
(2)
The crystals may enter or leave the size range via
either growth or flow. If G is the growth rate and Q is
the volumetric flow rate then [11]:
(3)
upon rearrangement, this equation becomes:
(4)
taking the limit as ΔL tends to zero, we have:
(5)
If we let τ = V/Q, MSMPR model applies and
, and equation (5) becomes:
(6)
Integration of equation (6) gives the function of
cumulative population vs. length:
(7)
where no is the population density at L= 0 and is
assumed to represent the nuclei.
Thus
(8)
9
Life Science Journal 2014;11(6s)
http://www.lifesciencesite.com
The second kinetic parameter that can be deduced
from CSD is the nucleation rate. If the nucleation rate,
Bo, is the rate appearance of near zero-sized particles,
an expression can be derived
(9)
Therefore, a semi logarithmic plot of the
population density vs. size yields no from the intercept
and G from the slope as illustrated in eqn. (8).
Since residence time (τ) is fixed, by the choice of
the experimental conditions, the nucleation and growth
rates can be determined simultaneously by a single
measurement of the CSD at steady state conditions.
were washed using a solution consisting of 50%
ethanol and 50% saturated KCl solution. The crystals
were then put in the drier overnight and 70oC and
allowed to cool to ambient temperature in a desiccator.
The dried crystals were subjected to sieve analysis
using a stack of sieves. The sieve sizes were 1400,
1000, 710, 500, 355, 250, 180, 125, 90, 75m and a
pan.
The crystals at each size cut were weighed and
subjected to chemical analysis using Pye Unicam Sp8
atomic
absorption
spectrophotometer.
The
concentration of the most important impurities (NaCl,
MgCl2, and CaCl2) were calculated after measuring the
concentrations of (Na+, Mg++).
3. Material and Methods
Apparatus
A cooled mixed suspension mixed product
removal crystallizer (MSMPR) was used. The
crystallizer is a jacketed vessel of one liter operating
volume. It is constructed from stainless steel. The
suspension inside the crystallizer is mixed using a 3blade-turbine impeller driven by an electric motor.
Four baffles are installed to enhance the mixing.
Cooling water is circulated through the jacket to keep
the crystallization temperature at the desired value. The
feed to the crystallizer is supplied from a constant level
tank of four liters capacity. This tank is fed from a feed
solution tank of about 70 liters capacity and is
constructed from Plexiglas. Continuous mixing is
maintained inside the two tanks by stirrers driven by
electric motors. Figure 1 shows a schematic diagram of
the crystallization system.
Experimental Method
The saturated feed solution at 40oC was prepared
by dissolving the required amount of KCl salt in the
feed solution tank. The solution was filtered using a
whatman 41 filter paper to remove the excess salt and
any other undissolved solid inerts. The filtered solution
was then poured into the feed solution tank and its
temperature was kept at 5oC above the saturation
temperature to prevent crystallization through pipes.
When the desired temperature of the feed solution was
reached, the constant level tank was filled with feed
solution by opening the valve between the feed solution
tank and the constant level tank. The temperature of the
solution in the constant level tank was the same as that
in the feed solution tank. The crystallizer was filled
with the solution via the valve between the constant
level tank and the crystallizer itself. The stirrer and the
cooling water pump were switched on. The flow rate
was measured by measuring cylinder and a stopwatch
and adjusted to the required value by the valve between
the constant level tank and the crystallizer. As the
crystallization temperature was achieved, the
crystallizer was run for ten residence times. After that,
the crystallizer output solution was filtered directly
using a whatman 42 filter paper. The collected crystals
4. Results and Discussion
Commercial grade KCl salt (wt. % KCl, wt. %
NaCl, Wt. % CaCl2) produced by the APC was used to
prepare the feed solution for all the experiments
performed in this study. Two different sets of
experiments were conducted: the first one aimed to
study the effect of the level of supersaturation on the
purity of the product, while the second one aimed to
study the effect of the residence time. The wt. % of
KCl salt and that of other impurities were plotted
against either the degree of supersaturation or the
residence time. A semi log plot of population density
(n) versus crystal size L for each one of the
experiments was used to determine the kinetic
parameters. Finally, the growth and nucleation rates
were plotted against the degree of supersaturation and
the residence time.
Effect of the level of supersaturation on the purity
of the crystalline product:
Six experiments at six different levels of
supersaturation (T=3, 6, 9, 12, 15 and 18oC) at a
constant residence time (=10min.) were performed to
study the effect of this parameter. Figure 2 shows that
as the degree of supersaturation, T, increases, the wt.
% of KCl, NaCl and MgCl2 in the crystalline product
decreases; while that of CaCl2 increases. The solubility
data given in Tables 1 and 2 show that the solubility of
CaCl2  KCl  MgCl2  NaCl. Moreover, Table 1
shows that the solubility of CaCl2 in the temperature
range (20-40oC), which is the range of concern in this
study, increases more dramatically than the others.
Therefore, as the level of supersaturation T increases,
more amount of CaCl2 crystals will be produced
leading to fewer amounts of other salts since the
amount of Cl- ions present in the feed solution which
are needed to produce the crystals of other salts is
constant as the feed solution is prepared to be saturated
at 40oC.
It can be deduced that the presence of Ca+2 ions in
the solution acts as crystallization inhibitor for
crystallization of KCl, NaCl and MgCl2.
10
Life Science Journal 2014;11(6s)
http://www.lifesciencesite.com
Effect of supersaturation on the crystallization
kinetic parameters:
Figure 3 shows that the growth rate decreases
while the nucleation rate increases as the level of
supersaturation increases. The magma density M
increases with the level of supersaturation. Its increase
is very dramatic for T  9o C as shown in Table 3.
This leads to high attrition between the suspended
crystals and between the crystals and solid surfaces in
the system leading to high rate of secondary nucleation
and to low rate of crystal growth due to crystal
breakage. Dash et. al. mentioned that for identical
residence times, higher suspension densities lead to
higher secondary nucleation rates and lower growth
rates [1].
Effect of residence time on the purity of the
crystalline product:
Six experiments at six different residence times
(=2.5, 5, 7.5, 10, 17.5 and 15min) and at constant level
of supersaturation (T=12oC) were performed to study
the effect of this parameter. Figure 4 shows that as 
increases, the wt.% of KCl decreases, those of NaCl
and CaCl2 increase; while that of MgCl2 is nearly
independent. It seems that the positively charged ions
suffer from high competition for the chloride ions
which are present in nearly constant amounts in the
feed solution. Table 4 shows the ionic radii of the ionic
species present in the crystallization solution. It is
known that, for a given identical charge, a smaller ion
would produce a more intense electric field. A stronger
field implies stronger ion-ion interactions and bonding,
therefore Na+, Ca+2 ions will attract the chloride ion
more than the K+ ion. This means that less amount of
Cl- ions will be available for K+ ions to produce KCl
salt. Increasing the residence time increases the
possibility of the reactions between Na+, Ca+2 and Clto reach equilibrium and hence to increase the
consumption of Cl- ions and in effect leading to less
amount of KCl salt.
Effect of the residence time on the crystallization
kinetic parameters:
Figure 5 shows that both the nucleation and
growth rates decrease as the residence time increases.
This was also obtained by Dash et. al. [1]. It is believed
that increasing the residence time increases the
dissolution rate due to the high mobility of the ions and
hence leading to lower growth and nucleation rates.
Comparison between the experimental kinetic
parameters and the predicted ones:
Equations 10 and 11 show the correlation
proposed by Dash et. al. [1] for the growth rate and that
given by Qian e.t al. [12] for the nucleation rate, and
they are used in the present study for comparison.
G = 5.1 x 10-6  -0.61
(10)
Bo = 3.58x108  -1089
(11)
Table 5 shows the experimental values of both G
and Bo correlated from data shown in Figure 6, and the
predicted values for both G and Bo calculated from
Equations 10 and 11. Accordingly, there is
compatibility in the general trend between the
experimental and predicted values of G and Bo.
Figure 6 shows a strong independence
relationship between the population density of the
crystals and the residence time. On the other hand,
Figure 7 shows that the population density of the
crystals does not depend on the level of
supersaturation, T.
Temperature, oC
Solubility g KCl/ 100 g H2O
T, oC
MT, g crystal/g H2O
Ionic Species
Ionic Radius(nm)
Table (1): Solubility of NaCl, MgCl2, and CaCl2. [12]
Temperature, oC
0.0
10.0
20.0
30.0
40.0
60.0
Table (2): Solubility of KCl [13]
20
24.95
30.65
34.218
35.599
37.298
Salt Solubility g salt/ 100g H2O
NaCl
MgCl2
CaCl2
35.6
52.8
59.5
35.7
53.5
65.0
35.8
54.5
74.5
36.1
56.0
102.0
36.4
57.5
37.1
61.0
137.0
34.85
38.506
Table (3): The variation of magma density with supersaturation
3
6
9
12
15
0.0042
0.0074
0.0054
0.0317
0.0330
Table (4): Ionic radii of different ionic species [14]
K+
Na+
Ca+2
1.33
0.95
0.99
11
Mg+2
0.65
41.15
40.386
18
0.0392
Cl1.81
Life Science Journal 2014;11(6s)
http://www.lifesciencesite.com
Table (5): Experimental and Predicted values of kinetic parameters
τ [min]
Experimental
Predicted
2.5min (150s)
G=5.111x10-7 (m/s)
G=2.39x10-7
Bo=4.198x104(no./S.L
Bo=20761x104
-7
5min (300s)
G=2.48x10
G=1.57x10-7
o
4
B =1.345x10
Bo=0.745x104
-77.5min (450s)
G=1.919x10
G=1.22x10-7
o
3
B =5.356x10
Bo=3.462x103
-7
10min (600s)
G=1.556x10
G=1.03x10-7
o
3
B =4.192x10
Bo=2.01x103
-7
12.5min (750s)
G=1.26x10
G=0.89x10-7
o
3
B =1.911x10
Bo=1.32x103
-7
15min (900s)
G=0.983x10
G=0.8x10-7
o
2
B =6.315x10
Bo=9034x102
Constant Level Tank
Over Flow
Cooling Water In
Cooling Water Out
Product Out
Crystallizer
Feed Solution Tank
Figure 1. Schematic diagram of the crystallization system
(a)
(b)
Figure 2. (a) Wt. % of KCl vs. degree of supersaturation, (b) Wt. % of CaCl2, MgCl2, NaCl vs. degree of
supersaturation
12
Life Science Journal 2014;11(6s)
Figure 3. (a)
http://www.lifesciencesite.com
Nucleation rate vs. degree of supersaturation, (b) Growth rate vs. degree of supersaturation
(a)
(b)
Figure 4. (a) Wt% of NaCl salt vs. residence Time, (b )Wt% of MgCl2, CaCl2, NaCl salts vs. residence time
(a)
(b)
Figure 5. (a) Nucleation rate vs. Residence Time, (b) Growth rate vs. residence time
13
Life Science Journal 2014;11(6s)
http://www.lifesciencesite.com
Figure 6. Population density vs. average crystal size at different degree of super saturation
Figure 7. Population density vs. average crystal size at different residence time
5. Conclusion
In this study, it has been shown that the ionic
properties of the impurity species have significant
effect on the purity of the crystalline product and on
the kinetic parameters of the crystallization. The
purity of KCl salt produced is affected adversely as
the level of supersaturation increases as well as the
residence time also increases. It has been noticed that
the nucleation rate increases while the growth rate
decreases as the level of supersaturation was
increased. Both rates were found to decrease as the
residence time was increased.
Nomenclature
Bo : nucleation rate (no./L. s)
G : crystal growth rate ( m/s)
L : characteristic crystal size (µm)
MT : magma density ( g crystal/ g water)
N : no. of crystals of size L and smaller in the magma
n : crystal population density function ( no. / L.µm)
no : crystal population density function at L=0
Q : volumetric flow rate (L/s)
t : time (s)
14
Life Science Journal 2014;11(6s)
http://www.lifesciencesite.com
ΔT: temperature difference between feed tank and
crystallizer (oC)
V : volume of mother liquid in magma (L)
τ : residence time (s)
6.
7.
Corresponding Author:
Dr. Khaled M. RAWAJFEH,
Chemical Engineering Department,
University of Jordan
Amman 11942,
Jordan
Phone : +962-6-5355000
Email: [email protected]
8.
9.
References
1. Dash S. R., S. Rohani and S. Sokhansanj,
‘Impurity Effect on Crystallization of KCl from
Aqueous Solutions in a MSMPR Crystallizer’.
The Canadian Journal of Chemical Engineering,
74, 265-278, (1996).
2. Bostaris G. D., Mason E. A., and Reid R. C.,
‘Incorporation of ionic impurities in crystal
growing from solution. The case of lead ions in
Potassium Chloride crystals’. AIChE Journal
13(4), 764-768, (1967)
3. Sarig S., Clasner A., Epestein J., A., and
Eidelman, Journal of Crystal Growth, 39, 255266, (1977)
4. Mullin J. W. Crystallization’, 2nd edition,
Butterworths, London, (1972)
5. Sayed S. A., and Larsen M. A., ‘ Effects of
NaCl and MgCl2 on the Vacuum Crystallization
of KCl’, Presented at the 75th annual meeting of
AIChE, Los Angeles, (1982)
10.
11.
12.
13.
14.
15.
3/23/2014
15
Koenig A., and Emons H. H., ‘Effect of
Electrolytic Impurities on the Crystallization of
Potassium Chloride’, Freiberg Forschurgh,
A683, 46-67, (1983)
Koenig A., Emons H. H., and Nyvlt J., The
Influence of Sodium Chloride on the driving
force of the crystallization of Potassium
Chloride from Aqueous Solution at 25oC’,
Cryst. Residence Technol., 22, 13-19, (1987)
Akl M., Nassar M. and Sayed S. A.. “influence
of impurities and suspension density on the
crystallization kinetics of Potassium Chloride”,
Chem. Eng. Tech., 63, 935-939, (1991)
Rohani S. and NG B., ‘Batch Crystallization of
KCl in the presence of soluble and Insoluble
Impurities’, Chemical Engineering Science, 47
(2), 367-379, (1992)
McCAbe W. L., and Smith J. C., ‘ Unit
Operations of Chemical Engineering’, 3rd
edition, McGraw-Hill, USA, (1976)
Myerson A. S., ‘Handbook of Industrial
Crystallization’, Butterworths, London, (1993)
Mullin J. W. Crystallization’, 2nd edition,
Butterworths, London, (1993)
Farelo F., Barchel G. V., and Offermann H.
‘Solid-Liquid equilibrium in the ternary system
NaCl-KCl-H2O’. The Canadian Journal of
Chemical Engineering, 71, 141-146, (1993)
Maron, S. and Prutton, C, Principles of physical
chemistry, 4th ed., U.S.A., (1969), 429. 62
Qian R. Y., Fang X. S., and Wang Z. K.,
‘Supersaturation and Crystallization Kinetics of
Potassium Chloride’, Ind. Eng. Chem. Res., 28,
844-850, (1989).